Objective To investigate the applicability of the Chinese version of the Edinburgh postnatal depression scale (EPDS) for postpartum depression screening in puerperal women using Rasch analysis, so as to provide reference for the clinical use of appropriate postpartum depression screening tools. Methods A cross-sectional study was conducted on 234 puerperal women in the outpatient department of obstetrics of a tertiary general hospital in Yangzhou, Jiangsu province. Rasch model was used to analyze the unidimensionality, reliability, difficulty and model fit of EPDS items. Results The Chinese version of EPDS had 10 items, and the Cronbach’s α coefficient was 0.812; it had unidimensionality, the reliability of the scale was 0.82, and the point biserial were 0.46 to 0.76. The fitting index of each item of the scale was good, the infit mean square value ranged from 0.64 to 1.29, and the outfit mean square value ranged from 0.62 to 1.34. The range of item difficulty was—2.26 to 3.16, matching well with the depression level of the participants. The options of each item were independent and had good discrimination. Conclusion The Chinese version of EPDS has good reliability and unidimensionality, and has good applicability in screening postpartum depression in puerperal women. Based on Rasch model analysis, it is found that the scale is primarily suitable for postpartum depression patients with mild to moderate severity, being less sensitive to high levels of depression. The effectiveness and sensitivity of the scale in clinical application can be further improved by adding items with higher difficulty levels.